The Hidden Edge in Markets No One Is Trading ft. Tom Babbedge

[00:00:00] Tom

So, we put it in, and I think it was one of our best performing markets, the next couple of years, and it’s continued to be a pretty good trender over the history we’ve been trading. Actually, for us, we feel that’s a bit too financialized. So, that’s not one we trade, but 2024, after 20 years of no trend, that was probably the biggest commodity trend there was, so.

[00:00:28] Intro

Imagine spending an hour with the world’s greatest traders. Imagine learning from their experiences, their successes and their failures. Imagine no more. Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world so you can take your manager due diligence or investment career to the next level.

Before we begin today’s conversation, remember to keep two things in mind. All the discussion we’ll have about investment performance is about the past and past performance does not guarantee or even infer anything about future performance. Also understand that there’s a significant risk of financial loss with all investment strategies and you need to request and understand the specific risks from the investment manager about their products before you make investment decisions. Here’s your host, veteran hedge fund manager Niels Kaastrup-Larsen.

[00:01:18] Niels

Welcome to another episode of the Open Interest series on Top Traders Unplugged, hosted by Moritz Seibert. In life as well as in trading, maintaining a spirit of curiosity and open-mindedness is key and this is precisely what the Open Interest series is all about.

Join Moritz as he engages in candid conversations with seasoned professionals from around the globe to uncover their insights, successes and failures, offering you a unique perspective on the investment landscape. So, with no further ado, please enjoy the conversation.

[00:01:54] Moritz

Hello and welcome to another episode of the Open Interest series on Top Traders Unplugged. This is episode number 21, and I’m your host, Moritz Siebert.

Today I’ll be speaking with Tom Babbedge from Gresham Quant, London, which is a unit of Gresham Investment Management LLC, a firm with a 35 year history in the commodity markets and more than US$8.1 billion of assets under management. Gresham Quant was founded in 2016 and specializes in the systematic trading of alternative and niche commodity markets, a topic which I find very interesting and I hope you will too.

I should note that this is the second time I’m inviting someone from Gresham to the show. The first time was with Scott Kerson, about 2.5 years ago, in December of 2023. Scott started Gresham Quant in 2016 and launched the ACAR program in 2017. ACA is short for Alternative Commodity Absolute Return Program, and it has one of the longest life track records in the alternative markets trend following space. I should also mention that Scott decided to leave Gresham Quant last year and that Tom is now in charge of ACAR, including all the research and trading processes.

Today, Tom and I will speak about Gresham’s portfolio, some of the markets they trade, why and how they decide to add markets, but also reasons for removing markets they previously traded. We’ll speak about their execution framework, trading costs, bid offer spreads, return dispersion, and much more.

But before we start, let me give you some background information on Tom. Tom is the chief Scientist at Gresham Investment Management and in charge of the ACAR program. Prior to Joining Gresham, from 2007 to 2016, Tom was a senior researcher at Winton and the personal researcher for Winton’s founder, David Harding. Before Winton, Tom was a postdoctoral researcher in the astrophysics group at Imperial College in London, working on galaxy evolution modeling and observations.

Tom holds a PhD in astrophysics from Imperial College, a Master of Physics from Bristol University, and an interesting anecdote, next to him having four kids, two cats, one puppy and 10 chickens at home, is that he helped Brian May, the guitarist of Queen, to complete his astrophysics PhD.

Wow, that’s a lot, Tom. I think it’s a long enough intro. I’ll stop here and I want to welcome you to the Open Interest series and thank you for joining me on the podcast today.

[00:04:11] Tom

Yeah, well, thank you, Moritz, longtime list, first encore, I guess, as they say. So, yeah, very excited to be here.

[00:04:18] Moritz

The ACAR program trades alternative commodity markets only. So, this is a key difference, I think, to other alternative markets, CTAs, you’re not the only one trading alternative markets systematically. But as far as I know, you’re the only one that only trades commodities. So, you’re not trading exotic interest rate swaps, or currencies, or any of the other financial, you know, niche markets, really only the commodities.

What would you say? Why is this? Why did you decide to go down that route and kind of like, you know, miss on the diversification opportunities and benefits that you could potentially get from these other markets?

[00:04:58] Tom

Yeah, and it almost sounds like a puzzle because I think the traditional, certainly in quant finances, you know, the only free lunch is diversification. If you’ve got 100 markets, you should trade 200. If you’ve got 200, you should trade 400. That’s the kind of arms race is how long is your market list. But I think there are two things there that I think are worth touching on. One is just incrementally adding one market to your 301st market really makes no difference for diversification. And that’s really because you’ve also got to think about what allocation you can give to those markets.

And if your 301st market has got, you know, 0.1% allocation or something, it really makes no difference. So, that you need a certain number of markets. But I think it’s also important what the markets are, because if you’ve got five-year treasury notes, and ten-year treasury notes, and two-year treasury notes, most of the time that’s kind of one market, not three, it’s not three degrees of freedom.

And I think, really, commodities are the one place in markets where you can really point to real structural reasons why you understand that they’re different. You can almost be reliant on them being different, even during risk-on/risk-off macro shocks; that they’ll be pretty resilient in taking their own course.

So, for example, when we trade South African sunflower seeds and US east coast power and, I don’t know, Chinese bitumen, maybe in a month or two period you might mathematically measure a high correlation or a low correlation. That almost doesn’t matter because you know, fundamentally, they’re entirely different.

They’re in different regions, they are addressing different needs, they have different consumers, hedges, producers, different drivers. You’ve got to ship something over the world or you’ve got to grow a crop over here. They’re just fundamentally different from a sort of molecular sense, a temporal sense, you know, a geographical sense. And that imbues the portfolio with a real inherent low correlation and high diversification.

And I think that’s less the case when you start considering more financialized markets like FX indices, fixed income broadly. You know, certainly if you look at indices most of the time they kind of look like one point something markets, not twenty different indices. And that’s particularly the case when there’s kind of macro shock. Everything collapses to one in terms of its correlation.

And if you restrict yourself to commodities, and particularly less financialized commodities, you’re kind of sidestepping that whole collapse-to-one effect. Your portfolio tends to be pretty resilient in terms of both internal diversification and its sort of low correlation to other things.

So, that’s one aspect, is diversification, there is actually as much, I would argue, even more diversification within the commodity space than within the wider universe. Now, the problem or the key issue here is to be able to express that diversification, you have to be able to meaningfully allocate across those different commodities. And so that’s where the question of how much money you are managing is an important one.

Because if you are managing billions and billions, then maybe you have got 150 commodities, but you’re going to have to put most of that allocation into the biggest and most liquid. So, for example, crude oil or Comex gold or maybe Dutch gas, if you’re sort of going down the liquidity a little bit. And that means that the allocation or the impact that you get from South African sunflower seeds, or Chinese bitumen, or US east coast power, is suddenly reduced. And what you’ll find is that unless you have a real strong capacity discipline, your portfolio won’t look diversified.

So, it’s kind of true. Just trading commodities on its own maybe won’t give you a very diversified portfolio unless you marry it to a capacity discipline. Once you do that, you can unlock accessing a whole range of very different and very diverse risk factors.

So, for example, we trade biofuels, we trade uranium. You’re tapping into the green transition, you’re tapping into nuclear power resurgence and how you power AI data centers. You’re tapping into all these different things; freight routes and the fact that globalization is shrinking into a sort of onshoring and regionalization, and you’ve got tariff wars and various political spheres that are sort of turning their backs on each other. That starts to make lots of things look very different in different parts of the world. And the supply and demand and the logistics all changes.

But the only reason you can tap into that is if you decide you’re not going to manage US$10 billion, because then, necessarily, you’re going to look like a sort of standard commodity allocation within a broader portfolio which is reasonably diversified but isn’t really able to stand on its own feet, I’d argue. So, it’s a choice.

[00:10:17] Moritz

Got it. Speaking about the capacity discipline which you mentioned, what would you think is the capacity of your program and the way you trade it today?

[00:10:26] Tom

Yeah, so, certainly on paper, during COVID, when commodity markets were at super high prices, super high volatilities, you’re thinking about units of risk that you can deploy, dollar risk. On paper, capacity in our markets could have been US$4 or US$5 billion and it wouldn’t have compromised diversification.

Now, the key thing is there, standing during COVID, did we think going forward that that capacity would necessarily stay at those levels or would it return to lower levels? And you could have filled up the coffers, as it were, and taken on US$4 or US$5 billion. We decided to half close. We were around a billion at that point.

The key for us is that existing clients can still benefit from organic performance growth within the portfolio over the next few years. Because, you know, the investment horizon that we think is relevant for a trending strategy is not months or even a year, it’s several years. And so, you want to have enough headroom, in terms of your capacity, that existing clients can benefit from that and you’re not giving them back the money just because you made 15%, 20% in a year.

So, for us, topline capacity is of the order of, let’s say, a couple of billion. But we would be closing before that because we want to maintain a headroom for growth within the portfolio. We don’t want to penalize current clients by taking on new money.

[00:11:53] Moritz

It’s quite a high number. I would have thought that the number might even be less than a billion. But I guess it also then means that you’re trading quite a lot of markets so that you can deploy, you know, risk of, say, a hundred… I’m not sure how many. You’ll tell me, in a second, if it’s 100, 150 or 200 markets. With that number, you can probably deploy risk and get to that US$1 billion or US$2 billion capacity number that you mentioned.

[00:12:18] Tom

Yeah, so we trade in the order of 150 markets and that’s been a number we bumped around for the last few years. I mean, it’s worth saying, the markets within that list has changed quite a lot over time. We’re quite active, in terms of being strict, on what maintains in the portfolio and what we bring in. So, for example, last year we dropped 44 markets and we added 31. So, that’s a reasonable change there. And that’s really because of our view of what’s alternative and what provides the best opportunity.

[00:12:48] Moritz

Let’s speak about this a bit more because that is really interesting. It’s quite, you know, when you say you’ve added 31 and you removed 44 and the base portfolio number was like 150. It’s kind of like, you know, you’re changing a third of your portfolio, something like that, or, you know, 25%. How do you form these decisions? How do you decide what goes in and what goes out?

[00:13:14] Tom

So, the first thing I would say, it’s not based on the back test. So, it’s not that you play with all the markets, you find the best ones that did the best results over the years and gives you the highest Sharpe in your backtest portfolio. That is the key to future misery, both for you and for investors.

[00:13:31] Moritz

Right.

[00:13:31] Tom

Because you’re selling them a Sharpe 2, and it turns out it’s a Sharpe sub 1, or something. It’s the classic issue in systematic finance because it’s so cheap and easy now to run your processes and explore parameter space.

So, for us, it’s all about the markets themselves. What are they? Who’s trading them? Why are they trading them? Where are they trading them on the curve? How does that market sit within the wider portfolio in terms of its relation to other things? So, if you think about, there’s certainly a cluster within our portfolio that I would call furnace. So, things like iron ore would cluster somewhat with things like flat glass panels and coke and coal, and that’s because they all kind of related to this sort of furnace concept, which is a real physical thing.

[00:14:18] Moritz

You’ve just mentioned, you figure out who’s trading them and where they’re trading them on the curve. How do you obtain that information and that sort of knowledge?

[00:14:28] Tom

Yeah, so, in terms of how do we evaluate a market and either drop it or add it? So, there are a number of metrics we have which, for us, we would label as their level of alternativeness. So, that can be as simple as how correlated are I to the big mainstream commodities, like things you find in BCOM?

And, obviously, we want to be diversified and different to Main Street commodities, so we would like lower correlation. So, for example, if there’s an expensive oil spread that looks alternative from a low liquidity perspective, that’s not interesting to us if it’s going to be highly correlated to mainstream crude. It’s just an expensive way to gain crude exposure. So, for us, liquidity isn’t the metric for alternativeness.

I think the other metric that often has been raised in the past is kind of barriers to entry. Is it hard to trade? You’ve got to have the right broker relationships. You know, maybe you can only access it through a swap. You know, for example, Chinese commodities in the past.

Again, for us, that’s not a criteria. You know, often that’s an outcome of the sort of markets we’re trading. But it’s not a qualifier, if you like. For us it’s how many speculators are in this market.

So, you can look at commitment to trader data. You can look at the sort of the level of open interest that goes to delivery versus gets sort of recycled back in. You can talk to brokers, you can ask who is trading this. For example, you know, last year, towards the end of last year I went to a biofuels conference and also container freight conference. Now no one else there was a quantitative trader. They were the standard people who want to hedge that risk in those markets or are producing in those markets.

If I found that every other of our, sort of, compatriot CTAs was at these conferences, that would not be another market for me. This is starting to become less interesting to us because it’s become a financialized mainstream sort of market. And we can measure these things as well. This isn’t just, I guess, qualitative.

So, there’s a great paper that my colleagues Adam Peddle and Yoav Git wrote recently, it was called You Trend, I’ll Follow: You Lead, I Follow. And that’s decomposing trend into two components. There’s a drift component, which I’d call kind of like the quantity of trend, how far a market moves. And there’s an auto-correlation component, which I would label as the quality of the trend because, clearly, a big shock up is not something useful to trend, but something that expresses 10% over a few months is a great trend for us.

So, there are these two components, and you can measure those for markets, and you can understand those, and you can see that they’re quite clearly linked to, for example, the level of speculative activity in that market. And you can see that difference, that relationship over the past sort of 30 years.

And there’s a quite clear separation in these properties for what we would term, an alternative commodity to more mainstream commodities. And that’s been persistent over multiple decades. And that’s an outcome of who’s trading it and why. And on that autocorrelation point, where you trade on the curve is another important point.

So, if you think about the crude curve, which goes out 5, 6, 7, 8, 9 years, the front of that curve is a financialized, speculator dominated, kind of noise news driven market. If you go five years out, that’s where the hedgers, and the consumers, and producers, are putting on the hedges. It’s a very different market in terms of who you’re trading and how that market is moving. And that autocorrelation, that quality of how the trend expresses itself out there, it’s almost like the difference between sort of news and weather at the front of the curve and kind of more like a climate effect further out the curve.

So, the trends may be similar, but they’re much more easily captured by a trend follower because they are expressed over a smoother and longer period. So, it’s those two components, the quantity of the trend and the quality of the trend. And what you find is in the markets that we kind of select for, those outcomes are both better.

[00:18:43] Moritz

I think that ties also into the paper that I think you wrote together with Scott, a couple of years ago, where you speak about trends in alternative markets or that the trends haven’t gone away, they’ve just moved to a different neighborhood. And I think one of the key arguments of that paper is that you say that these alternative markets have better and longer-lasting trends.

Is that something that you have checked up on? Do you still agree with that? Would you say that is generically true, that these more niche markets have better trending properties?

[00:19:15] Tom

Yeah, well, you can break that down into the trend quality and the trend quantity aspects, and you can measure them for mainstream market commodities and alternative commodities for 30 odd years, and you can see an air gap between them on those. And I think that’s really driven by, if you think about something like a coal mine, it takes decades perhaps to get permission to start a new coal mine. And they don’t turn them off overnight just if the cost of coal sort of drops below their break even. They’ll keep running them at a loss for a number of years because shutting down the coal mine is hugely expensive. Once you have done that, it’s very hard to open it up again.

And then when you dig up that coal, you’ve got to ship it somewhere over the globe. That takes time. Storage of coal, you can do it for a bit, but it starts to degrade, especially if it’s outside. So, there’s an in inelasticity in these sort of markets between supply and demand.

And when there’s that difference, and it can’t be balanced, that the only way it balances is that the price moves a long way in one direction or the other. Either attracting new suppliers or persuading people to switch to some alternative. And that’s what, inherently, I think for us, you can point to in commodities there’s a reason that drives trend.

I think it’s less obvious in some other markets what is driving the trend. But here you can sort of see there’s a structural inherent reason for that and that gives you the ability to select markets based on those properties rather than rely on whether, in a backtest, the trend was a good Sharpe or not.

So, an example of that is, back in 2016, when Scott and I were building the nascent ACAR sort of model and considering which markets to include at launch, we were looking at European carbon, which is, at the time, really the only carbon market around with any liquidity. At that point, I think it was, maybe, eight years of history or something, and the eight-year trend history of doing trend on that market was, I think, negative, in terms of Sharpe it was negative. So, perhaps we should have not considered it because, from that selection, just trend doesn’t work.

Now we put it in because, when we looked at our metrics of what we think makes an alternative market and that has the opportunity to and potential for large sustained trends, you know, it ticked those boxes. So, we put it in.

I think it was one of our best performing markets the next couple of years, and it’s continued to be a pretty good trender over the history we’ve been trading. I think people would probably say something similar about cocoa as well, but actually for us we feel that’s a bit too financialized. So, it’s not one we trade. But 2024, after 20 years of no trend, that was probably the biggest commodity trend there was.

[00:22:03] Moritz

Interesting. Before we go more into the markets, I’d like a question on your program. When you put positions on do you treat all the markets in the same way in terms of risk budgeting, the longs and the shorts, is it all symmetric in your program?

[00:22:19] Tom

In terms of longs and shorts in in futures land, yeah, we would view that as a symmetric problem. Obviously, there are a few edge cases because these are commodities and there’s kind of real-world boundaries in some cases.

Clearly we don’t trade near the front of the curve in oil but technically all went negative during COVID. I think a more real-life example of that for us, last year actually, is the Californian carbon scheme. There’s an effective auction floor price, each set, each year, and that sort of staggers up and the idea is it sort of building the price over time. I think it was April last year that the price in the futures market actually came down quite close to that floor.

Now, if we were short, clearly that market, then, is not naturally able to trend shorter because there’s an effective buffer there that’s sort of preventing you. So, in that case, we overrode that model so it couldn’t increase its positions. But it was free to sort of organically reduce the position if it wished to, if the trend went the other way, which it did eventually and the price sort of was able to move back to a region where it could move in either direction freely. But no, in general, for futures, long and short are very similar. There’s not really any difference. There’s someone on each side of that.

[00:23:38] Moritz

Right. And also you’re putting the same risk on, say, a long propane position as you would on a long freight position.

[00:23:46] Tom

Yeah, I mean there is an overall risk allocation to each market. Now, that will be determined in the optimization process based on things like correlation to the rest of the book and diversification, also the liquidity of that market. Our largest position in the market is one of the drivers of our sort of top-end view on capacity across the book.

So, we want to be able to get out of a max conviction, long or short, within say a week or so in a stress scenario. So that’s one of the drivers. But it’s tying that to our requirement for the high diversification and ability to allocate across the markets. When you put those together, that’s what gives you the top line capacity.

So, that will mean there is a range of weights, I guess, to markets. And, for example, some of our smaller, newer markets, we will add them in with a smaller weight because you can spend a lot of time getting lots of quotes from brokers before you actually start trading a market. But, you know, the acid test is always actually going out and getting those fills done.

And if you’re in a low liquidity market, it’s best to err on the side of caution there and initially sort of start small and then build into a market once you’ve gained that sort of evidence on the ground of how it trades.

[00:25:11] Moritz

Do you trade most or maybe even all of your markets through the big bank brokerage houses, or do you require and depend on a lot of specialist OTC brokers?

[00:25:23] Tom

Yeah, this is one of the downsides of the choices we’ve made, which is that broadly you cannot plug your computer into the exchange and just electronically trade. You have to do it through getting quotes and the old fashioned way of effectively getting on the phone and talking to a network of dealers and brokers. Some of that will be done by banks. But for a lot of our markets there are specialist brokers, and they will be getting the best price.

And in fact, if you were to ask a bank, you’ll find that often behind the scenes they’re then going to the specialist brokers. So, you don’t want that middleman, but you do want competition on pricing. So, we have a very wide network of brokers. And when we bring on a new market, it may well be that we have to onboard some new broker.

We’ve done that in Japan in the past where we had to get them actually signed up with US regulation so they could trade with a non-Japanese entity. But that then gave us the right broker relationships to get the liquidity and the good quotes that we wanted. So that’s a big part of the job. The execution desk handle is that broker network relationship because it’s really key for the markets we trade.

[00:26:32] Moritz

Do you then, I mean, the markets you trade, are they all cleared as a futures contract through an exchange clearinghouse, or do you have OTC counterparty risk as a result of your relationship with brokers?

[00:26:46] Tom

They’re all futures markets, so they’re all cleared on the exchange. So, there isn’t that counterparty aspect. We do some of those, for example, Chinese commodity futures we do via swap. But that money is all kept offshore.

[00:27:02] Moritz

Right, so, you would have the counterparty exposure or the exposure to the swap counterparty and the underlying. The swap then references the futures contract. The futures is clear through an exchange but your risk is still to the swap counterparty where you exchange margin.

[00:27:18] Tom

Yes, well, in that case, even there, we actually keep it with a third party, the end of margining. So, it’s outside of the hands of the swap counterparty.

[00:27:33] Moritz

Now one of the really interesting things you mentioned, Tom, during our wrap call is the execution process that you’re using, which is far less systematic than most CTAs or trend following funds would trade.

I mean, if I understood it correctly, I’d say it’s a very discretionary execution process that, you know, spans over a couple of days. Could you explain how that works?

[00:28:02] Tom

Yeah, yeah. So, my colleague, Yoav Git, he did write a blog on this, about a year or so ago, called The Waiting Game. And to give you the backdrop, as a systematic CTA, execution is often just electronic and really the execution desk there is really just monitoring that execution.

And typically, you’ll have some smart algo that has been given some window. Maybe it’s two days, or two hours, or maybe it’s half a day, or maybe it’s just an hour to get that fill done; you know, 100 lots of gold, please, or whatever it is. And when you benchmark that against a TWAP or a VWAP, you know, it all looks fine, slippage hits the models, everything’s in line.

Now, what CTA’s found over the past 10, 15 years was, as they became bigger and started to manage multi billions, that slippage bill starts to rise. As you know, a typical liquid CTA, maybe they’re spending, I don’t know, say, 70 basis points a year, maybe even a percent on slippage transaction costs.

And the solution was to actually cut that up into different parts of the day. So, rather than the model waking up and wanting to trade 100 lots of gold, it wakes up with a fifth of the allocation and says, oh, my model, at 10 AM, wants to trade 20. Then it will wake up at noon and that will want to also buy. And then it wakes up at 2 PM, in the afternoon, and that also wants to buy.

Overall, those five models kind of, probably, want to buy about 100 lots, but each one is kind of benchmarked to its own new window. And so, internally, they all look great, and the slippage within them is all kind of same as before. But the problem is, that is really an autocorrelated trade over the whole day, and it’s kind of now hidden there. And it kind of looks like the slippage is in control, but it’s actually risen.

And the other thing to think about is, if you think about the risk that you’re taking on a daily basis and in your positions in a portfolio, if you’re targeting 15% risk annualized, that’s roughly 1% variance a day, right? That’s your kind of risk budget. And nearly all of that is just the positions you’re holding. It’s not the trades you’re doing on the day. Because if your turnover is once a month, 20 days in a month, so that’s something like probably a twentieth of your daily variance, maybe 5% of that 1% daily variance is in the trades. So, it’s a small amount.

And then if you’re chopping that up and trying to do it within a shorter period, and you’re doing it all across the day, the effective risk you’re seeing from the trading desk is even lower. And the trades are pretty uncorrelated across markets. Positions might be correlated, but trades are uncorrelated.

So again, that diversification means the effective risk you are seeing in the portfolio from trading, on a given day, is actually very small. It’s less than a percentage of that 1%. It’s very small. And that means even if you have really fancy, super-fast price prediction algos that can really get you the best price, it’s kind of second order, third order in terms of impact on the book.

Now, the thing that we are doing differently, and this is something that Yoav first started on our sister strategy SAFI (which is a sort of financial instrument strategy, several years ago, with great success, and it’s something we’ve rolled out to ACAR at the beginning of last year), is to turn that question around, actually, and say we’re going to give the desk a higher level of effective risk allocation, if you like. So, they have, in a sense, an allocation to the trading desk and you put risk controls around that and tracking. That’s important, you want to be able to track their variance versus the model.

But what it means is a couple of things. You’re giving them more risk, so, it’s more important. By giving them the risk and letting them choose over what time period they’re going to do that trade, and now they might even take several days, they’re effectively the thing that’s controlling the cost in the book.

So, the traditional way to reduce costs in a CTA is to turn down the speed of your trading to the signals. So, it’s just slower signals, you’re doing less trades. This way you can actually speed up the signals. Because now you’ve got this sort of intelligent buffer, which is the trading desk, which can sort of in an intelligent manner decide to either slow down or just act on the model. And there’s a reason that’s nice, is because normally fast signals are better before cost, right? So faster the better in terms of signal space.

But when you get to the execution, actually being more patient can really pay off. Because if you’re being passive and you’re being a liquidity provider, you’re not the person who is crossing the spread and paying that slippage, someone else is. So, if you can wait for that market to come to you and half the time it will, on average, because markets are pretty random. You’re not crossing the spread.

Now that doesn’t make a huge difference in a very liquid market because spreads might be a few basis points, but we trade freight, and freight spreads can be easily 100 basis points. So, that’s a really meaningful difference.

And what we found is, if you’ve built the right risk framework around the effective risk allocation you’ve given the desks (they don’t have an infinite allocation, you know, they have bounds), so, they can wait for a price for a certain amount of time. But if the tracking error from where we want to be has risen too much, is too much risk in that unrealized trade, then they’ll have to trade it. But that freedom and that passivity means that suddenly what was a very large slippage bill can potentially be very small.

And that was very interesting last year. We’ve been running it for, now, just over a year. Our slippage bill last year was effectively close to zero compared to previous years, in a relative sense. And that’s partly because you’re not crossing the spread.

But the other interesting effect is, if you woke up today and wanted to do 100 lots of, I don’t know, some biodiesel, and the market was moving away from you, you didn’t trade it. When the model wakes up the next day, maybe it only wanted 80 lots. So, you’ve already just saved having to trade 20 lots. So, there’s a kind of invisible slippage saving that you make as well because, inevitably, the model will change its ideal position slightly each day.

So, you get these two effects. You can kind of view it as the trading desk being an anti-correlated strategy to trend. And so, actually, what you also find is that you can increase the risk, the risk target, not the risk target, but the risk scalar that you’re running the model at because you’re getting this extra diversification effect. So, it’s been a really interesting result, and yeah, quite different to the normal way a CTA thinks about execution, which is just get it done, essentially.

[00:35:31] Moritz

In your example with the biofuel contracts, like 100 lots, and then the next day it’s 80 lots, is that because of a daily vol control, or risk control framework, targeting framework that you’re running where you’re adjusting position sizes, potentially daily?

[00:35:48] Tom

Yeah, so, I don’t think this is anything other than perhaps the standard CTA approach. But broadly your position is proportional to the strength of your signal, which might come from a number of different trend horizons, but it’s inversely proportional to the volatility of the market because you’re taking a risk adjusted view on positioning. So yeah, if risk was spiking up, then you would need less of that increasing trade the following day.

[00:36:13] Moritz

Right, right. I’m asking not because I want to go down the volatility targeting route, but in the example you mentioned, I mean, to give you the counterexample, possibly the next day the model could wake up and say, I don’t want to have 100 lots of the biofuels contract that you’re trading, I want to have 125 lots, so I want more. And you haven’t executed the trade yet and the price has run away from you.

You’ve kind of like, you’ve missed one day. At some point, you know, there is a chance or a possibility that you’ll run out of time. The market trades away from you. You want to buy more lots. And at the end of that time span, you’re kind of like forced to not only cross the spread, because you want to have the position, but you’re also buying at a much higher price than you would have on day one. How often did that happen?

[00:37:01] Tom

Exactly. That that happens as well. And that’s the key to making sure you’ve got the right monitoring and, I guess, risk allocation to the desk.

So, yes, it went from 100 lots. Nothing was done. In 220 lots, nothing was done. They’re starting to get closer and closer to their sort of risk allocation in terms of difference they can be from the portfolio; the variance that’s in the trade that hasn’t been done. At some point they’ll reach that, and then they’ll just fall back onto, I guess, the normal way you do trading, which is you go out and do the trade.

[00:37:34] Moritz

Right.

[00:37:36] Tom

And yeah, so what you’ll find is the variance in your slippage is higher because sometimes you’ll get unlucky. The market keeps moving away from you until you’re forced to do the trade. But if you think about sort of markets being close to random, it’s pretty close, 50/50 every day, whether a market is going up or down. And so, you can almost say that half the time you’re not crossing the spread. And that’s a pretty big saving locked in.

Sometimes the market will move away from you and you’ll lose out on getting into the trend for a bit until you’re forced to get into it. Equally, if the market goes towards you and goes past you, you’re actually making positive, you’re making money on your slippage, if you like. It’s negative slippage. And you see that happening too.

Certainly, I mean, for us, Liberation Day wasn’t really an event. We were up for the month, but certainly we saw very big propane moves. And the way we were executing the model, wanting to cut its positioning in that sort of more passive manner, saved us a lot on slippage that would have burnt us in terms of the market moving away and then coming back. And we could see that there.

So, that’s anecdotal because both those things can happen. But, on average, you get those two effects of the market moving a long way past you in either direction. One of those hurts you and one of these helps you.

But in the middle you’re not crossing the spread anymore and that’s a kind of guaranteed part of it. So, it’s noisier. But once you’ve run that experiment for a year or so, you start to see there’s quite a clear benefit to it.

[00:39:17] Moritz

Probably fair to summarize this as, you have the systematic trade engine and then a nonsystematic discretionary alpha generator based on execution that sits on top of that systematic trade engine.

[00:39:31] Tom

I’d hesitate to go quite that far. Or at least I would argue that anyone who’s going out and doing their trading, therefore, arguably is doing something discretionary in how they’re doing the trading. It’s just that the approach we’ve taken is to be more passive. And there are very clear controls and monitoring around how passive that becomes or how long you take but it’s just a different implementation.

You could almost write that down as a trading execution algorithm, if you liked. Just like the fasted algos are encoded. And for me that doesn’t mean it’s really turned it into a discretionary aspect. It’s just giving a budget to the trading desk to wait for the markets.

[00:40:20] Moritz

Right, to allow them to be more passive and in a waiting mode, and waiting for the market to come to you.

[00:40:26] Tom

Yes, it’s less that the trader wakes up that day and thinks, fine, he wants to buy biodiesel, but I’ve got a really bad feeling about biodiesel, and I had read some Twitter account saying it’s going to go the other way, and I’m going to make my own bet. You’re not trading against the model, you’ve just got a bigger window in which to carry out that. And the choice of when you carry it out is really driven by how the market is moving either towards or away from you.

So, it’s not really discretionary in what I would call a capital D. It’s still part of our systematic framework and fits within that.

[00:41:09] Moritz

Changing gears a little bit. I’d like to speak about, if you can, I mean, some of the conversations you have with clients or the experience you have when you speak with clients. Why do they give money to Gresham Quant? Is it because they want an alternative market specialist? Is it because you’re so uncorrelated? Is it because you’re commodities only? What are the main drivers and key decision points for clients when they invest with you?

[00:41:35] Tom

Yeah, so certainly, if a prospective client comes to us and really they’re dipping their toe into trend for the first time, and they’re wanting something that is kind of what I would call core trend sort of convexity to equity shocks, it’s very clear that’s not what we’re going to be offering you. Right? You should go and find a cheap, management fee only, liquid trend that will go up when stock markets go down and it won’t cost you much. That’s your core allocation.

Where I think we fit is that we are applying the similar technology to capture similar trend moves and positive convexity and positive skew, but we’re doing it in different risk factors. And that means that, for example, our correlation to The SocGen CTA is 20%, 25%. So, we are a new diversifying allocation within your portfolio.

Now you might be thinking you want that from a diversification perspective, or you might be looking at it as simply an uncorrelated alpha stream. I think either of those are valid. But I think one thing that does resonate quite strongly with clients is the fact that, because it’s commodities and there’s a very clear reason why they should trend, I think it shortcuts a lot of the kind of faith aspect that you almost have to put into any system or any signal.

Because if someone has come up with a really complex machine learning signal, and the back test is great, and maybe it’s had two or three years of great performance, there’s still a lack of clarity on where that alpha is coming from, exactly. And therefore, is it going to be persistent and in what environments is it going to deliver to you?

And I can be pretty sure, if it then delivers you three years of zero to negative Sharpe, people are not going to keep it in their portfolio because they don’t understand, one, why it was working, and two, why it’s no longer working.

Now, the pain of trend is that it’s not a super high Sharpe and you can get periods of long drawdown and certainly, you know, recently we’ve been in a drawdown of several years. Now, the key resilience there is understanding that in trend, and certainly in commodity trend, it’s almost an inherent outcome of those markets inelastic supply and demand and the diversification is very concrete.

And that means that you can still believe and understand that it has a long run positive Sharpe, and you’re in one of the periods where unfortunately it’s sampling the left side of that distribution, if you like. That’s one aspect. I think another one is concerns around inflation.

So, if you think about commodities, there’s almost a chicken and egg situation here about when you measure inflation, you’re measuring prices of goods and goods are commodities. If inflation’s going up, it’s really the commodities going up. They’re kind of two sides of the same coin. And there’s a very strong and clear positive convexity for commodity trend and inflation and deflation.

In fact, there was a recent paper we put out called The Eras Tour and it’s just quite interesting how, if you add commodity trend, even basic commodity trend, to something like the standard 60/40 portfolio, it’s just very clearly a much better protection for you during either high inflation, or low inflation, or rising inflation sort of periods.

And so, I think those are the kind of aspects. I think a lot of clients have found that they like alternative market CTAs and they probably already have one, but they kind of feel that the commodity aspect is underweighted and that might be because of a capacity question around how much you’re managing, how much you can put into the capacity sub portfolio. And we’re almost a top up there, right? We’re topping up that commodity side.

So yeah, it varies by client and I think it also varies by, I guess, how much experience people have had or trend and where it fits in the portfolio. But broadly it’s low correlation. It’s what I would hope is long-term alpha, not guaranteed every single year, and it’s a clear narrative of why it works, and when it works, and that link to inflation as well.

[00:46:26] Moritz

Well, let’s bring this conversation to a close, Tom. Thanks again for coming onto the podcast today. It was really interesting, and for our listeners, as usual, I’ll put the most important takeaways of my chat with Tom into our show notes. And should you have any questions, please reach out to us and send us an email. You can contact us at info@toptradersunplugged.com.

Thank you for listening and until next time on the Open Interest series.

[00:46:52] Ending

Thanks for listening to top Traders Unplugged. If you feel you learned something of value from today’s episode. The best way to stay updated is to go on over to iTunes and subscribe to the show so that you’ll be sure to get all the new episodes as they’re released. We have some amazing guests lined up for you, and to ensure our show continues to grow, please leave us an honest rating and review and iTunes. It only takes a minute and it’s the best way to show us you love the podcast. We’ll see you next time on Top Traders Unplugged.